Deepfakes, Democracy and the 3-Hour Takedown Rule

On February 10, 2026, India’s Ministry of Electronics and Information Technology rolled out sweeping new amendments to its IT Rules, fundamentally redrawing the country’s digital speech landscape.1 Platforms now face a sharply reduced threehour window, down from the previous thirty-six, to take down deepfakes or synthetic content once ordered by the government. There’s also a new requirement: all AI-generated content must be clearly labelled. Miss these obligations, or even stumble, and the consequences are severe: platforms lose their “safe harbour” protection, becoming as legally exposed as the users who post the content in question. The rationale for this change is obvious and deeply felt. Synthetic media has been weaponised in India: think election deepfakes, non-consensual intimate imagery, and state-led information warfare across borders. The harms are not abstract. They’re immediate, sometimes devastating, and threaten the very foundation of trust in public discourse. Yet, the government’s cure comes loaded with costs of its own. By bypassing judicial checks and procedural safeguards, the system places extraordinary power in the hands of the executive, delegating censorship to private intermediaries, and, in doing so, operating outside the judicial oversight and proportionality that India’s Constitution demands. This article contends that the 2026 amendment directly undermines Article 19 of the Constitution and the standards set in Shreya Singhal v Union of India,2 converting once-neutral intermediaries into de facto enforcement arms for the state. By comparing India’s model to regulatory frameworks in the EU and the United States, this article argues that New Delhi’s solution, intended to defend democracy, places democracy itself in peril by making truth and lies equally disposable at the stroke of an executive order.